Optical fiber networks are playing an increasingly important role in high-speed communications networks. To handle the increasing data rates and bandwidths expected of these networks, advances in optical communication technology continue to be made at a brisk pace. One of the vital components of optical fiber networks that continues to be improved upon is the optical amplifier, which is charged with the task of initially transmitting, or later boosting, an optical signal along an optical fiber. Such optical signals may be characterized as either single-channel or multiple-channel signals. Multiple-channel signals are normally implemented via wavelength division multiplexing (WDM), whereby the wavelength spectrum of the optical signal is apportioned into multiple wavelength segments, each of which implements a separate communications channel, thereby increasing overall communications bandwidth. As optical amplifiers for these optical networks become more sophisticated, test systems used to measure the performance of the optical amplifiers necessarily must be improved in order to provide the developers and manufacturers of the amplifiers with the parametric information they require to thoroughly characterize their products.
A particularly important characteristic of an optical amplifier is its polarization dependent gain (PDG). Designers and manufacturers of optical amplifiers normally seek to minimize changes in the intensity level of the output signal of these devices when the polarization of the incoming optical signal changes state. Generally speaking, the measured PDG of an optical amplifier is its maximum output signal minus its minimum output signal over substantially all optical polarization states.
Several different methods for analyzing the intensity level of the output signal of an optical amplifier under varying input conditions currently exist. FIG. 1 shows a first optical amplifier test system 100 for testing an optical amplifier 1. A saturating optical signal source 110 generates an optical signal that is capable of saturating the input of the optical amplifier 1. The saturating optical signal source 110 may be either a single-channel or multiple-channel WDM source. At the saturation level, changes in intensity level of the output signal of the optical amplifier 1 are minimal when the intensity level of the input signal varies about the saturation level. Use of the saturating optical signal source 110 thus eliminates a potential cause of change in the output signal level when measuring the intensity level of the output signal of the optical amplifier 1. The normal operating condition of the optical amplifier assumes such a saturated mode.
The optical signal generated by the saturating optical signal source 110 is then quickly polarized in a random fashion by a fast polarization scrambler 120. The terms “fast” and “quick” refer to the fact that the polarization states of the optical signal are cycled through more quickly than the recovery/relaxation rate of the optical amplifier 1 will allow the amplifier 1 to substantially react to the changes in polarization state. In other words, while the polarization states are being changed, the intensity level of the output signal of the optical amplifier 1 remains essentially constant. Normally, the fast polarization scrambler randomly cycles through some small subset of possible polarization states well distributed throughout the possible states represented by a Poincaré sphere.
The resulting polarized signal from the fast polarization scrambler 120 is then passed as an input signal to the optical amplifier 1. The resulting output signal of the optical amplifier 1, operating in saturation, is then analyzed, typically by way of an optical signal analyzer 140, which can determine the intensity level and wavelength of the output signal. Due to the action of the fast polarization scrambler 120, the intensity level of the output signal for a single channel test (or for each WDM channel in the case of a multiple channel WDM optical signal source) is essentially constant throughout all polarization states, thus representing an average output value for the channel being tested.
For multiple channel optical amplifier tests, the intensity level of each WDM channel is often measured separately. Unfortunately, if each WDM channel exhibits the same polarization state, as is the case with the polarized signal from the fast polarization scrambler 120, the optical amplifier 1 is likely to exhibit the phenomenon of “polarization hole-burning.” This phenomenon results in a lower-than-normal output signal of the optical amplifier 1 for each of the WDM channels of the optical amplifier 1. Many (but not all) amplifier designers and manufacturers require that polarization hole-burning not affect the intensity level measurement for each channel. Thus, a second optical amplifier test system 200, as shown in FIG. 2, may employ a WDM polarization mixer 130 for the polarized signal of the fast polarization scrambler 120 to mix the polarization of the WDM channels with respect to each other to avoid polarization hole-burning. Thus, the WDM polarization scrambler 130 essentially eliminates the possible effects of polarization hole-burning from the intensity level measurement of each WDM channel. The resulting signal from the WDM polarization mixer 130 is then used as the input signal for the optical amplifier 1 under test.
While such a measurement is useful, the polarization dependent gain of a channel, which requires detection of minimum and maximum output signal levels across a variety of polarization states, cannot be determined effectively in this manner due to the speed at which the fast polarization scrambler 120 cycles through the polarization states.
To address this issue, FIG. 3 shows a third optical amplifier test system 300 that employs a slow polarization scrambler 150 in place of the fast polarization scrambler 120 of FIG. 1. The slow polarization scrambler 150 rotates through polarization states at a rate at which the relaxation/recovery rate of the optical amplifier 1 allows the output signal of the optical amplifier 1 to react essentially completely to the change in the polarization state of the input signal. One such slow polarization scrambler is the Agilent Technologies 11896A Polarization controller, which utilizes a series of optical fiber loops to accomplish any possible polarization state. Typically, within the third optical amplifier test system 300, the slow polarization scrambler 150 need cover only a small number of the possible elliptical polarization states in a random fashion in order to effectively detect the minimum and maximum output signal levels of the optical amplifier 1 for each channel tested. Typically, the slow polarization scrambler 150 can cycle through the used polarization states in approximately five seconds while still providing the polarization state coverage necessary to test optical amplifier 1 sufficiently. Also, the slow polarization scrambler 150 normally employs random cycling through elliptical polarization states while avoiding pure linear or circular polarization states, which also may cause the aforementioned polarization hole-burning. As described earlier, such a phenomenon would have a detrimental effect on determining the true PDG of the optical amplifier 1.
Also, as discussed before in relation to second optical amplifier test system 200, addition of a WDM polarization mixer 130 at the output of the slow polarization scrambler 150, as shown in a fourth optical amplifier test system 400 (FIG. 4), will essentially eliminate the effects of that phenomenon from the PDG measurements of multiple WDM channels.
Unfortunately, the third and fourth optical amplifier test systems 300, 400 ordinarily introduce another type of error in the PDG measurement of optical amplifier 1. This error is in the form of a polarization dependent loss (PDL) of the portion of the optical path between the output of the optical amplifier 1 and the optical spectrum analyzer 140. Many newer optical amplifiers operate as waveguides that provide a separate optical path for each WDM channel. This amplifier architecture causes the PDG of each channel of an optical amplifier 1 to potentially vary greatly from channel to channel. Each separate channel of such an amplifier may be required to exhibit a subjectively acceptable PDG, which is normally in the range of 0.3 dB or less. However, with a normal prior art test system PDL of about 0.12 dB, the PDG measurement of some channels of the optical amplifier, as measured by the optical signal analyzer 140, may be skewed significantly either up or down, possibly resulting in an unacceptably inaccurate determination of PDG for one or more channels of the optical amplifier 1.
Therefore, from the foregoing, a new optical amplifier test system and method that greatly reduces the polarization dependent loss of the test system, thus reducing the amount of error in the polarization dependent gain measurement of an optical amplifier, would be advantageous.